Moviedb_neo4j Chatbot

This project is designed to work with the provided Movie Database. Using a different database may lead to unexpected errors.

Prerequisites

• Install Miniconda and Conda

  • Download and install from Miniconda.

• Install Ollama and Hugging Face Model

  • Install Ollama from Ollama.
  • Search for lakkeo/stable-cypher-instruct-3b on Hugging Face.
  • Click Use this model and select Ollama.
  • Choose either Q5_K_M or Q8_0 (NOTE - Q8_0 less hallucinates and has higher accuracy and I would not recommend to go to higher model as it needs more compute).
  • Copy the command shown (e.g., ollama run hf.co/lakkeo/stable-cypher-instruct-3b:Q8_0).
  • Open the terminal and run the copied command.
  • The first installation may take some time, but subsequent runs will be faster depending on your hardware specs.

• Install Neo4j Desktop and Setup Database

  • Download the Neo4j Desktop App Image from the Deployment Center.
  • Select the Linux App Image and download it.
  • Navigate to the directory where the file is downloaded:

    cd /path/to/downloaded/file

  • Set the file as executable and run it:

    chmod +x <file_name>
    ./<file_name>

  • Paste the activation key when prompted.
  • Turn on the Movie Database.
  • Navigate to Apps and start the Browser.
  • Open the Neo4j Browser App.
  • Copy the contents of MoviesDataset.txt, paste them into the Neo4j Browser and Execute it to store those values in the databse.

• Install Git and Clone the Repository

  • Ensure Git is installed.
  • Clone the repository:

    git clone https://github.com/Jitmandal051004/neo4j_chatbot_py.git <repo_folder_name>
    cd <repo_folder>

  • After cloning you can remove the ./images folder

• Set Up the Conda Environment

  • Using Conda (Recommended)
    • Create and activate the environment:

      conda env create -f environment.yml
      conda activate <your_env_name>

  • Using Python Virtual Environment (venv)
    • Create and activate the virtual environment:

      python -m venv venv
      source venv/bin/activate 

    • Install dependencies:

      python install -r requirement.txt

  • Additional Notes:
    • Conda is recommended as it handles package dependencies more efficiently.
    • If using venv, ensure you have the correct Python version installed.

Running the Project

• Run the Python script:

  python text2cypher.py

• Start the Streamlit app:

  streamlit run QueryApp.py

• Open the browser and fill in the password.
• Write your query and execute it.

Example Queries

• "Show me all the movies available."
• "Who acted in The Matrix?"
• "Which movies were directed by Lana Wachowski?"
• "List all co-actors of Keanu Reeves."
• "Find me all movies released after the year 2000."
• "Show movies that feature both Keanu Reeves and Laurence Fishburne."
• "Who produced The Matrix?"
• "Find all actors born after 1970."
• "List movies with the tagline containing 'freedom'."
• "Get all movies sorted by release year."

Troubleshooting

• If a query does not return the expected results, click Submit again.
• The model may sometimes generate excessive text or hallucinate responses.
• This happens because large language models predict responses based on patterns in training data.
• If a query returns incorrect results, try rephrasing and submitting again.

ENJOY

📌 WebApp User Interface

📊 Graphical Visualization

📋 Tabular Representation

Q8_0 model result